INDEX

Explanations

mentions of the word "tar."

oai_token-act-pair · gpt-3.5-turbo

references to tar and related terms

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Blooming

-0.85

 Howell

-0.80

eva

-0.73

 Simmons

-0.70

Anderson

-0.70

 Saunders

-0.67

 Bulldogs

-0.65

 Goldberg

-0.64

ESA

-0.64

 Garden

-0.64

POSITIVE LOGITS

tar

4.20

tar

2.30

Tar

2.08

Tar

1.83

 oils

1.19

 Keystone

1.05

SHA

1.04

oil

0.97

 pulp

0.96

SHA

0.95

Activations Density 0.028%

mentions of the word "tar."

references to tar and related terms

No Comments

No Known Activations